Laracon US 2025 recap: What “Laravel Boost” means for your website

At Laracon US 2025, I was lucky enough to see the Laravel team preview Laravel Boost, which was then released in a public beta on 13 August. Boost helps AI assistants work inside a Laravel project in a safer and more accurate way, using Laravel’s own tools and guidance.

What is Laravel Boost?

Laravel Boost is a free, open-source package you add to a project. It runs inside the app and exposes a Laravel-aware MCP server with more than 15 safe tools. These tools let an AI assistant read facts like your Laravel version, routes, database schema, recent errors and logs. Boost also ships Laravel-maintained AI guidelines and a documentation API built from tens of thousands of versioned docs across the Laravel ecosystem.

MCP in plain English

MCP stands for Model Context Protocol. Think of it like a universal USB-C style plug for AI software. It standardises how AI assistants connect to approved data and tools so they can ask for information or perform limited actions with permissions you control. In practice, an MCP server is a small program that sits next to your app and answers specific questions from the assistant, for example “what routes exist” or “show the last 20 log lines”. This avoids custom, one-off integrations and improves safety, since the assistant only sees what the server is allowed to expose.

Why clients should care

Faster delivery with fewer wrong turns

Without context, AI can guess the wrong API or pattern. Boost closes that gap by letting the assistant consult the right, version-specific docs and check the real state of the app. That reduces back-and-forth on small features, tests and common fixes.

Better consistency across the codebase

Boost installs Laravel’s own AI guidelines so suggestions follow modern framework conventions and popular packages you may already use, such as Livewire, Inertia and Filament. That consistency helps long-term maintainability.

Safer debugging and clearer visibility

Because the tools are scoped, an assistant can read the last framework error, browser logs or database schema to propose a targeted fix, while we keep code review and testing in place. Laravel’s announcement stresses that all generated code should be treated as a draft and reviewed.

What this looks like in practice

• Small feature work: add a rate limit, adjust validation or generate a test. The assistant consults the docs that match your project versions and drafts code for us to refine.

• Bug triage: for a 500, the assistant can fetch the last error, scan recent browser logs, spot a typo and propose a fix for review.

• Documentation with context: instead of generic search results, the assistant queries Laravel’s documentation API that Boost provides, aligned to the versions in your app.

Security and privacy

Boost runs inside the development environment. The MCP server exposes only the specific tools you install, which means access is narrow and auditable. We decide what is enabled and we keep our normal safeguards: tests, peer review and staging before release.

Where this fits in our workflow

• Pragmatic use: Boost is ideal for routine tasks and investigation, not architectural decisions.

• Governance: every AI-assisted change is reviewed by a developer and covered by tests where appropriate. Laravel recommends keeping code review non-negotiable, which matches how we work.

• Cost: Boost is free under the MIT license.

The headline for your business

With Boost in the toolkit you can expect quicker turnaround on small changes, cleaner code that follows current Laravel standards, and faster diagnosis when something goes wrong. That means steadier progress on your roadmap and less time lost to avoidable mistakes.